What's new
What's new

Measure right angles better?

jscpm

Titanium
Joined
May 4, 2010
Location
Cambridge, MA
The standard way I have to measure right angles on blocks is that I stand the block up on a surface plate and touch a right angle standard to it, which is a cylinder. Then I shine a light and see if any light comes through.

This method leaves a lot to be desired. For one thing, it just measures the line of contact with the cylinder. Also, if you get light, it doesn't really tell you where the problem is, it just tells you that there is a problem.

Is there a better way to be doing this?
 
Squaremaster or differential levels if you need to measure high accuracy squares. If you have the time a autocollimator and a optical square.
 
I use three methods.

1) A "leaning" cylinder. Mine is a 6-in. cylinder square. You rotate it until you black out the gap, then read the calibrated marks on the cylinder.

2) A comparator square, which is an indicator on a stand with a "bumper". Needs a calibration source.

3) A magnetic cylinder square, but this only works if the work is flat.

In all cases you need a surface plate.
 

Attachments

  • P1060989.jpg
    P1060989.jpg
    90.1 KB · Views: 663
  • 9146 Comparator Square.jpg
    9146 Comparator Square.jpg
    25.2 KB · Views: 596
  • contentonly.jpg
    contentonly.jpg
    18.9 KB · Views: 554
Maybe off topic but if you need .0005" shim stock disassemble an old solid state capacitor from the outside of a car distributer. They look like tiny soup cans with a wire sticking out one end. They have a coil of metal (aluminum?) coiled up inside separated with an insulating plastic. It is about 5/8" wide and however much you want to pull off. If you need to grind within a thousandth it might do the trick in shimming.
 
Along the lines Spencer has already indicated, the low tech but quite accurate way is to have a surface plate and a cylinder square as references, and something like a Taft Pierce comparator square or Squarol. You can improvise the latter; mine by adding a hardened circular arc "bumper" to a heavy B&S indicator stand. This turned out roughly equivalent to the Taft item, at a fraction of the price.

https://www.subtool.com/tp/9146_taft-peirce_comparator_square.html
 
Should add that one of the first things to do after getting something like a Squarol setup together is map every square in the shop. I was happily surprised that most were in very good tune, but one of the "best" was way out of spec and a couple of combination square heads needed tuning (with a thin diamond file) to get back to square. I recheck any heavily used square each year or so -- should be more often in a commercial shop.
 
It's always about how accurate you need to be, this dictates the equipment/methods you'll use. For most work that has to be within .0005 over a given length what Spencer and Pete have suggested about using an arc segment are good enough for most work. By adding a segment of an arc to an ordinary indicator stand that's stout enough you make contact at the same relative point on the work as you sweep along the arc segment. If you have two parallel sides that you've indicated to verify that they are parallel (or even if they're not exactly parallel) you can split the difference as you grind them in. You're using the indicator for a comparative value, not a quantitative value in your readings. The cylinders, Squaremasters, and the other equipment sure is nice to use but can be expensive and make you reluctant to bring it out. I'll drag out mine now and then just to use it but for most work it's faster/easier for me to just swivel the indicator around to the other side of the stand where the arc segment is permanently mounted and use whatever length of rod gets me the proper height. The arc segment doesn't even have to be exactly ground, just a smooth and continuous surface will do since you're only using one spot on it. That's what most of the die makers I worked with used unless they had money to burn on a cylinder square or the other devices. Try it.
 
Last edited:
Not trying to :stirthepot: but holding something up against and using "light" etc. isn't measuring. To measure then a tangible result is needed, not an approving nod.
 
The standard way I have to measure right angles on blocks is that I stand the block up on a surface plate and touch a right angle standard to it, which is a cylinder. Then I shine a light and see if any light comes through.

This method leaves a lot to be desired. For one thing, it just measures the line of contact with the cylinder. Also, if you get light, it doesn't really tell you where the problem is, it just tells you that there is a problem.

Is there a better way to be doing this?
.
use granite square which you can get up to 36x36" and is heavy. literally can hold surface gage and .00005" indicator on granite square and indicate part for squareness to surface plate.
.
thats how they measure big parts on a surface plate thats 12 by 16 feet. literally inspector standing kneeling on surface plate with protective booties over his shoes. surface plate got to weigh easily over 10 tons
 
another way.

using a surface gage that has the triangular notch in the bottom, place a ball bearing in the notch. (this does the same thing as the curved portion on the bottom of the suburban squaring feature.) then set your test indicator at the correct height so it is in line with the curve on the ball vertically and at whatever the height you wish to measure. use a square standard such as a master square or cylindrical square to set the indicator. you can then check your part to an actual value of perpendicularity at that height.
 
Maybe off topic but if you need .0005" shim stock disassemble an old solid state capacitor from the outside of a car distributer. They look like tiny soup cans with a wire sticking out one end. They have a coil of metal (aluminum?) coiled up inside separated with an insulating plastic. It is about 5/8" wide and however much you want to pull off. If you need to grind within a thousandth it might do the trick in shimming.

Aluminum foil is also .0005.
 
I think I'm repeating info here but I'll be succinct:

You must Quantify! Feeler gages with a cylindrical square are perfectly ok. Or an other method you must quantify.
 
another way.

using a surface gage that has the triangular notch in the bottom, place a ball bearing in the notch. (this does the same thing as the curved portion on the bottom of the suburban squaring feature.) then set your test indicator at the correct height so it is in line with the curve on the ball vertically and at whatever the height you wish to measure. use a square standard such as a master square or cylindrical square to set the indicator. you can then check your part to an actual value of perpendicularity at that height.

Lat scraping class Nathaniel (tool and die guy) showed me a neat trick. Most Starrett surface gages have a ball on the end of the upright rod. Flip it over so the ball end is in the notch, get the indicator on the horizontal arm. This setup will do in a pinch. Just in case if you got no spare balls handy.

dee
;-D
 
depending what you need, 1st chose the way to craft it, and maybe, if you are sure on what you do & the machine, than you don't even need to measure :)

i build a 200mm cube, on right angles, ( "vinclat" would be the local word for such precision )

finish on a grinding machine

measuring on a cnc mill table; this machine was inspected before :) i got less than 0.01 perpendicularity, and this is more than i needed :)
 








 
Back
Top